Signal processing techniques for cognitive radio networks

Abstract

In the last couple of decades, the telecommunications have experienced a drastic development. The increasing demand of added value services and the need of wide-band connections for supporting their diffusion is rapidly saturating the frequencies available for communications. With all the spectrum bands already exclusively allocated, it is becoming extremely hard to find unused frequency bands to deploy new services, or enhance the existing ones. The Cognitive Radio networks (CRNs) have been recently proposed as a possible solution to these inefficiency problems. CRNs introduce a more flexible and dynamic assignment of radio frequencies which promise to mitigate the ever increasing spectrum shortage problem. In a CRN there are two categories of user (and providers) usually referred to as the primary user (providers) and the secondary users (providers). The Primary User is the owner of the license for that frequency band and has always the right of using the radio resource. The Secondary User is an un-licensed user that can access the spectrum bands of a primary user without interfering with any primary communication. A CRN consists of two “sub-networks”: the primary network (PN), which is the network of primary transmitter and receivers, and the secondary network (SN), which conversely is the network of secondary receivers and transmitters. The PN has the license for accessing the spectrum bands, while the secondary networks does not. However, the SN can access the frequency bands left un-used by the PN. In this doctorate thesis, the spectrum scarcity problem is addressed in the scenarios of cognitive radio network, both at the primary and secondary network side. The goal of the reaches activity has been proposing and implementing innovative methods for improving the spectrum utilization, mitigating the ever-increasing spectrum scarcity problem, supporting the diffusion of CRNs and sustaining the coexistence of primary and secondary users. The research field dedicated to the primary networks has been mainly concentrated on 4G/5G systems for their inherit flexibility in managing the frequency resources. In such systems achieving and then maintaining time and frequency synchronization between the transmitter and the receiver is one the most critical issues. In this thesis two innovative synchronization methods for 4G networks are proposed, named the twin test and the rand test. In addition the Extended modified generalized Q function (EMGQ) are derived to express the system error probabilities in a closed form. One of the more important operation of a SN is identifying the under-utilized frequency bands that can be used for secondary communications (i.e. the spectrum holes). The process of detecing the primary users and identifying spectrum holes is the spectrum sensing. The decision about the presence or absence the signal transmitted by the primary user can be obtained by comparing a decision metric against a fixed threshold. In this thesis, many innovative spectrum sensing methods are proposed. The main advantage of the devised solutions is that they require no information about the signal of interest, without dramatically growing up the system complexity. All the proposed algorithms have been applied for detecting primary communications in the 4G and 5G systems. In a CRN, the possibility of trading the radio spectrum offers new business opportunities for both primary and secondary providers that share the radio resource according to fair regulations. Unfortunately, in this scenario an unfair user could hide its radio signal under the noise floor and avoid being charged for using the spectrum. This operating scenario is completely different from what has been studied in the literature about cognitive radio users. In this dissertation, the new scenario is described and a method for detecting the deceitful users hiding under the noise floor in the presence of fair user (ehiter primary or seconadary) is proposed.